Re: [blink-dev] Intent to ship: WebGPU: Dual source blending

2024-09-05 Thread 'Corentin Wallez' via blink-dev
Hey Alex, Thank you for your answer and the example of where TAG was able to bring alignment to various APIs. I think what makes the situation difficult here is that WebGPU is a very active specification that gains new minor features or developer visible improvements monthly, if not more often (w

Re: [blink-dev] Intent to ship: WebGPU: Dual source blending

2024-09-04 Thread Alex Russell
Hey Corentin, Thanks for the thoughtful note on TAG engagement. The idea here isn't that the TAG will provide deep guidance about your design regarding GPU features -- y'all are the GPU experts! -- the general pattern here is the TAG helps to nudge groups to align with patterns it is seeing ac

Re: [blink-dev] Intent to ship: WebGPU: Dual source blending

2024-09-02 Thread 'Corentin Wallez' via blink-dev
Hey all, Thank you for the LGTMs. *Finch flag* I think there is a misunderstanding about this Finch flag: the flag is meant to contain a list of comma-separated GPUFeatureNames that are disabled on the GPU process side (which is then reflected in Blink). So it is a general mechanism to disable a

Re: [blink-dev] Intent to ship: WebGPU: Dual source blending

2024-08-28 Thread Daniel Bratell
LGTM3 (after two LGTM1s from Alex and Yoav). The finch flag is just a followup on Domenic's question. It seems the issue is that the documented flag doesn't actually correspond to the on-off switch for this feature so please update the finch flag documentation to reflect the correct switch. Ju

Re: [blink-dev] Intent to ship: WebGPU: Dual source blending

2024-08-28 Thread Alex Russell
Hey folks, LGTM1 with nits. First, it would be good for y'all to FYI this to the TAG, even if you don't block on review. I agree it's small and would wave it through. I know Daniel has questions about the Finch flag name, so would be good to have those sorted before this goes out too. Thanks,

Re: [blink-dev] Intent to ship: WebGPU: Dual source blending

2024-08-28 Thread 'Corentin Wallez' via blink-dev
On Wed, Aug 28, 2024 at 12:36 PM Yoav Weiss (@Shopify) < yoavwe...@chromium.org> wrote: > LGTM1 > > On Fri, Aug 23, 2024 at 2:46 PM 'François Beaufort' via blink-dev < > blink-dev@chromium.org> wrote: > >> Adding +Corentin Wallez to the thread >> >> On Fri, Aug 23, 2024 at 2:29 PM François Beaufo

Re: [blink-dev] Intent to ship: WebGPU: Dual source blending

2024-08-28 Thread Yoav Weiss (@Shopify)
LGTM1 On Fri, Aug 23, 2024 at 2:46 PM 'François Beaufort' via blink-dev < blink-dev@chromium.org> wrote: > Adding +Corentin Wallez to the thread > > On Fri, Aug 23, 2024 at 2:29 PM François Beaufort > wrote: > >> >> >> On Fri, Aug 23, 2024 at 7:01 AM Domenic Denicola >> wrote: >> >>> >>> >>> O

Re: [blink-dev] Intent to ship: WebGPU: Dual source blending

2024-08-23 Thread 'François Beaufort' via blink-dev
Adding +Corentin Wallez to the thread On Fri, Aug 23, 2024 at 2:29 PM François Beaufort wrote: > > > On Fri, Aug 23, 2024 at 7:01 AM Domenic Denicola > wrote: > >> >> >> On Thu, Aug 22, 2024 at 9:13 PM 'François Beaufort' via blink-dev < >> blink-dev@chromium.org> wrote: >> >>> Contact emails

Re: [blink-dev] Intent to ship: WebGPU: Dual source blending

2024-08-23 Thread 'François Beaufort' via blink-dev
On Fri, Aug 23, 2024 at 7:01 AM Domenic Denicola wrote: > > > On Thu, Aug 22, 2024 at 9:13 PM 'François Beaufort' via blink-dev < > blink-dev@chromium.org> wrote: > >> Contact emails >> >> fbeauf...@google.com >> >> Explainer >> >> The “dual-source-blending” GPU feature adds additional blend fact

Re: [blink-dev] Intent to ship: WebGPU: Dual source blending

2024-08-22 Thread Domenic Denicola
On Thu, Aug 22, 2024 at 9:13 PM 'François Beaufort' via blink-dev < blink-dev@chromium.org> wrote: > Contact emails > > fbeauf...@google.com > > Explainer > > The “dual-source-blending” GPU feature adds additional blend factors and > the WGSL @blend_src attribute to allow a fragment shader to blen